Institut für Geodäsie und Geoinformation der Universität Bonn - PDF

Description
Institut für Geodäsie und Geoinformation der Universität Bonn Wet Path Delay Corrections from Line-of-Sight Observations of Effelsberg s Water Vapour Radiometer for Geodetic VLBI Sessions Inaugural Dissertation

Please download to get full document.

View again

of 82
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
Information
Category:

Presentations & Public Speaking

Publish on:

Views: 36 | Pages: 82

Extension: PDF | Download: 0

Share
Transcript
Institut für Geodäsie und Geoinformation der Universität Bonn Wet Path Delay Corrections from Line-of-Sight Observations of Effelsberg s Water Vapour Radiometer for Geodetic VLBI Sessions Inaugural Dissertation zur Erlangung des Grades Doktor-Ingenieur (Dr.-Ing.) der Landwirtschaftlichen Fakultät der Rheinischen Friedrich-Wilhelms-Universität Bonn vorgelegt am 31. Januar 2012 von M.Sc. Jungho Cho aus Seoul Referent: Koreferenten: Priv.-Doz. Dr.-Ing. Axel Nothnagel Prof. Dr.-Ing. Heiner Kuhlmann Assoc. Prof. Dipl.-Ing. Dr. techn. Johannes Böhm Tag der mündlichen Prüfung: 28. Juni 2012 Publikation: Diese Dissertation ist auf dem Hochschulschriftenserver der ULB Bonn elektronisch publiziert. Erscheinungsjahr: 2012 Zusammenfassung Wasserdampfinduzierte Refraktionseffekte der elektromagnetischen Wellen stellen die zurzeit größte Fehlerquelle bei Messverfahren der Satellitengeodäsie, wie z.b. GPS und VLBI, dar. Die Problematik rührt hauptsächlich her von der stark variierenden Verteilung von atmosphärischem Wasserdampf sowohl in der Zeit als auch im Raum. Im Allgemeinen können diese Laufzeitverzögerungen durch den feuchten Anteil der Atmosphäre nicht exakt genug durch atmosphärische Modelle berechnet werden, die herkömmlich in Satellitengeodäsieanwendungen genutzt werden. In den vergangenen Jahrzehnten hat die Wasserdampfradiometrie ein großes Potential entwickelt, um den atmosphärischen Wasserdampfbestandteil zu messen. Allerdings ist der Prozess der Umrechnung von gemessenen Helligkeitstemperaturen in Laufzeitverzögerungen stark von gleichzeitig durchgeführten Radiosondenmessungen abhängig. Dabei werden die Messergebnisse von an aufsteigenden Ballons befestigten Wettersensoren für verschiedene Druckstufen per Radiosignal ausgesendet. Leider werden periodische Radiosondenbeobachtungen aber nur selten in der Nähe des Wasserdampfradiometers (WVR) durchgeführt. Dem gegenüber besteht seit einigen Jahren die Möglichkeit, ein numerisches Wettermodell anstelle der Radiosondenergebnisse zu nutzen. Ein numerisches Wettermodell kann meteorologische Profile für solche Orte liefern, wo eine Radiosonde nicht verfügbar ist. Der Schwerpunkt dieser Dissertation liegt hauptsächlich auf der verbesserten Bestimmung der Laufzeitverzögerungen durch den feuchten Anteil der Atmosphäre in der geodätischen VLBI, wobei die Wasserdampfradiometerbeobachtungen am Radioteleskop in Effelsberg genutzt werden. Verglichen mit anderen Wasserdampfradiometern hat dieses Instrument große Vorteile hinsichtlich der Messwertgewinnung. Es zeigt immer in dieselbe Richtung wie die VLBI-Antenne, weil es im Primärfokus des Teleskopes installiert ist. In oder in der Nähe von Effelsberg werden jedoch keine Radiosondenbeobachtungen durchgeführt. Um diese Schwäche zu beheben, wurde ein numerisches Wettermodell des European Centre for Medium Range Weather Forecasts (ECMWF) für die Bestimmung von Kalibrierwerten herangezogen. Es liefert für das Radioteleskop in Effelsberg meteorologische Daten wie z.b. Druck, Temperatur und Wasserdampfdruck. Solche Profile wurden in einem Strahlungsübertragungsmodell verarbeitet, welches theoretische Messungen der Helligkeitstemperatur ermittelt und diese in Laufzeitverzögerungen durch den feuchten Anteil der Atmosphäre umwandelt. Um die Laufzeitverzögerungen durch den feuchten Anteil der Atmosphäre aus Wasserdampfradiometermessungen und die Modelle besser vergleichen zu können, wurden alle Laufzeitverzögerungen durch den feuchten Anteil der Atmosphäre auf die Zenitrichtung (Zenith Wet Delays, ZWD) bezogen. Der Vergleich hatte zum Ergebnis, dass die ZWDs der Modelle einen um ca. 30 mm höheren Wert zeigten als jene, die mit einem Wasserdampfradiometer gemessen wurden. Im Vergleich zu GPS-abgeleiteten ZWDs betrugen die durchschnittlichen Offsets der Modelle und des Wasserdampfradiometers -4.3±11.0 mm beziehungsweise -44.8±24.0 mm. Diese ZWD- Vergleiche haben gezeigt, dass eine Korrektur der WVR ZWDs erforderlich ist. Außerdem hatte es den Anschein, dass die rohen WVR-ZWD-Messungen geglättet werden sollten, um das Rauschen des Instruments zu reduzieren. Für die Fehlerkorrektur wurden außerdem in jeder einzelnen Session durchschnittliche Offsets zwischen den Modellen und den Wasserdampfradiometern berechnet und angesetzt. Allerdings zeigte sich schon hier, dass die interne Kalibrierung des Instruments einige Defizite aufwies und die Ergebnisse dadurch in ihrer Genauigkeit eingeschränkt waren. Die Korrekturen an den Laufzeitverzögerungen in Zenitrichtung aus verschiedenen Ansätzen wurden in fünf geodätischen VLBI-Sessionen verwendet und die Auswirkungen auf die Basislinienwiederholbarkeit und Höhengenauigkeit untersucht. Es stellte sich heraus, dass die Basislinienwiederholbarkeit bei manchen Basislinien verbessert werden konnte, wenn Offsets an den gemessenen WVR-Ergebnissen angebracht wurden. Die Verbesserung war jedoch kleiner als 1 Prozent. Obwohl die Höhengenauigkeit, ausgedrückt als Root Mean Squared Error (RMS) und Weighted RMS (WRMS), um den Faktor 2 verbessert werden konnte, zeigte die Höhenkomponente selbst eine größere Ablage von den Ursprungswerten als erwartet. Als Ursache dafür wurde die Vielzahl der zu schätzenden Parameter und ihre zum Teil hohen Korrelationen identifiziert. Die Schlussfolgerung dieser Untersuchung ist somit, dass die Waserdampfradiometerbeobachtungen in Effelsberg noch nicht gänzlich für die Fehlerbehebung der Laufzeitverzögerungen durch den feuchten Anteil der Atmosphäre geeignet sind, was hauptsächlich auf die Unvollkommenheit einer instrumentellen Kalibrierung zurückzuführen ist. Es werden weitere Studien mit einer größeren Zahl von WVR- Messwerten mit einer verbesserten Kalibrierung des WVR notwendig sein, um die Zweckmäßigkeit des Wasserdampfradiometers für die Fehlerbehebung der Laufzeitverzögerungen durch den feuchten Anteil der Atmosphäre in der geodätischen VLBI abschließend nachweisen zu können. Summary Water vapour induced excess path lengths in electromagnetic waves have been one of the most unmanageable errors in space geodesy, such as GPS and VLBI. The difficulty mainly comes from the highly variable distribution of atmospheric water vapour both in time and space. In general, these wet path delays cannot be estimated accurately by atmospheric models that are conventionally used in space geodetic applications. In the last few decades, water vapour radiometry has shown great potential for measuring atmospheric water vapour content. However, the wet path delay retrieval processes are strongly dependent on radiosonde data, although periodic radiosonde observations are rarely available in the vicinity of water vapour radiometers (WVRs). Radiosonde observations are weather profiles from balloon starts which are transmitted by radio signals. On the other hand, the possibility of using a numerical weather model (NWM) instead of a radiosonde has been on the increase in recent years. NWM can provide meteorological profiles for those places where radiosonde data is not available. The focus of this thesis is mainly on the improvement of the wet path delay corrections in geodetic VLBI sessions using the WVR observations at the 100m Effelsberg radio telescope. Compared to other WVRs, the Effelsberg one has a great advantage in terms of observation. It always points at the same direction as the VLBI antenna because it has been installed on the prime focus cabin of the telescope. However the Effelsberg station does not make periodic radiosonde observations. To overcome this weakness, the numerical weather model of the European Centre of Medium Range Weather Forecasts (ECMWF) was introduced. It provides meteorological profiles over Effelsberg such as atmospheric pressures, temperatures, and water vapour pressures. Those profiles were processed by a radiative transfer model, which calculates theoretical measurements of brightness temperature and converts them into wet path delays. These two models were combined to be compared with WVRobserved wet path delays. For a better comparison between wet path delays from the WVR and the models, zenith wet delays (ZWDs) were used. As the results of the comparison illustrate, ZWDs from the models showed higher values than the WVR-measured ones by roughly 30 mm. For comparison with GPS-derived values, average offsets and standard deviations of the models and the WVR were -4.3±11.0 mm and -44.8±24.0 mm, respectively. From these ZWD comparisons it was found that further corrections to the WVR ZWDs are necessary. In addition, the noisy behaviour of the raw WVR ZWD measurements should be smoothed by a running mean method before application. In addition, averaged offsets between the models and the WVR measurements should be determined for the correction of individual sessions. However, already at this step it became obvious that the instrumental calibrations of the radiometer are far from being mature resulting in erroneous absorption profiles. ZWDs from the WVR measurements with different levels of corrections were applied as corrections to the wet components of the atmospheric refraction in the five geodetic VLBI sessions. Impacts on baseline repeatability and height precision by these were investigated. As the results show, the baseline repeatability was improved in terms of Root Mean Squared Error (RMS) when the offset correction was applied. However, the improvement was less than one percent. Although the repeatability of the height component was improved in terms of Weighted RMS (WRMS) with respect to the short term mean height by a factor of 2, the height component itself showed a larger deviation from the original value than that expected from the ZWD corrections. A possible reason is that the estimation of the many parameters in the least squares adjustment can easily affected the height parameter. The conclusion of this study is that the Effelsberg WVR observations are not perfectly suited for wet path delay corrections yet. This is mainly due to the imperfectness of instrumental calibration. Further studies based on an increased number of WVR data with better internal calibrations seems to be necessary to make a final judgment regarding the usefulness of the WVR for wet path delay corrections in geodetic VLBI. Contents 1. Introduction Motivation and objectives Motivation Objectives Water vapour in Space Geodesy Path delays induced by atmosphere Path delay induced by water vapour Path delay correction in geodetic VLBI Water vapour sensors Water Vapour Radiometry Brightness temperature Absorption Model Retrieval coefficients Various WVRs and their application for Geodesy Effelsberg WVR instrument co-located with VLBI Numerical Weather Model and Radiative Transfer Model ECMWF MonoRTM Profile mode Calculations Error propagation Wet Path Delay Comparisons and Readjustments Wet path delays from the WVR at Effelsberg Wet path delays from MonoRTM(ECMWF) Comparison with GPS-ZWDs readjustment of ZWDs ZWDs Applications to Geodetic VLBI and its Results Impacts on wet path delay corrections in geodetic VLBI Impacts on baseline repeatability Impacts on height precision Summary and Discussion Conclusions 79 Appendix 80 List of figures 96 List of tables 98 References 99 6 1. Introduction 1.1 Motivation and objectives Motivation Atmospheric water vapour degrades the accuracy of the results of space geodetic observations due to permanent electric dipole moments. It creates excess path lengths by retarding (slowing and bending) the propagation of the electromagnetic waves that are used in global positioning system (GPS) and very long baseline interferometry (VLBI) observations. It is known that the excess path lengths are less than 30~40 cm at the most, and are the primary obstacles of space geodesy because of the highly variable distribution of water vapour in the atmosphere. According to Askne and Nordius (1987), this wet path delay cannot be determined by only using surface meteorological data with an accuracy of 1 cm or better in the zenith direction. To cope with this deficit, geodetic VLBI analysts normally estimate the wet path delay contributions via various approaches. However, the number of unknown parameters increases considerably and the results still leave room for improvement for many space geodetic applications. It is known that improved accuracy can be achieved by using remote sensing techniques. Several authors (Resch et al., 1979, 1984; Ware et al., 1986, 1993; Kuehn et al., 1991, 1993; Johansson et al., 1993; Teitelbaum et al., 1996; Tahmoush and Rogers, 2000; Oswald et al., 2005; Nothnagel et al., 2007) suggested the use of water vapour radiometers (WVRs). A WVR measures the brightness temperature from the thermal emission of water molecules. In order to use a WVR for geodetic purposes, a conversion process from the brightness temperature to the wet path delay is necessary. Since Elgered et al. (1991) presented the usefulness of WVRs to geodetic VLBI, the quality of WVRs has constantly improved. However, wet path delay retrieval from the water vapour content has always been dependent on radiosonde data. Radiosonde weather profiles originate from weather balloons which transmit their data via radio signals. Only a few WVR stations have access to radiosonde data acquired in the immediate vicinity. With the advent of numerical weather models (NWM), a new method for the conversion of the brightness temperature or the readjustment of the wet path delay corrections may be possible, even for stations where radiosonde data is not available. Another modern development is that the observatory at Effelsberg has been operating a WVR with a new concept. Most WVRs measure brightness temperatures at only two distinct frequency channels. The Effelsberg WVR, however, possesses a receiver with 25 channels of 0.9 MHz bandwidth each, spanning from 18.3 to 26.0 GHz. Another promising feature of the WVR at Effelsberg is that it always points in the same direction as the VLBI antenna, continuously changing directions during geodetic VLBI sessions. In the absence of regular radiosonde observations in close vicinity to Effelsberg, it appears to be a suitable approach to use a NWM for improving the retrieval of the wet path delay corrections. 7 1.1.2 Objectives The primary hypothesis of this thesis is that the new concept of the Effelsberg WVR together with data from NWM yields improved results for the geodetic parameters estimated from VLBI observations. In contrast to conventional two-channel WVR, the Effelsberg WVR scans the water vapour emission spectrum on multi-channels. This sampling allows the separation of the emission from instrument effects and from the emission of cloud water (Tahmoush and Rogers, 2000). In addition, the WVR can always keep the line-of-sight direction of the VLBI antenna. In order to use the WVR for geodetic purposes, the conversion process between parameters in different units is indispensable. WVR-measured brightness temperatures in Kelvin can be converted into wet path delays in millimetres via the so-called the retrieval process. Every retrieval algorithm includes conversion coefficients. The coefficients are derived from the relationship between WVR-measured brightness temperatures and radiosonde-derived wet path delays. Unfortunately, Effelsberg has no regular radiosonde observations. To cope with this deficit, a NWM will be introduced to provide meteorological profiles over Effelsberg. In order to calculate brightness temperature measurements based on the profiles, a radiative transfer (RT) model will be introduced. At the microwave frequencies used in this study, the radiation from water vapour molecules is an integrated quantity of the two opposite processes that are the emission and the absorption, as depicted in figure 1-1. A series of interactions of the two processes along a line-of-sight are called the cascade process. Chandrasekhar (1960) presented the radiative transfer equation of energy transfer in electromagnetic waves to describe such a cascade process. In combination with NWM, RT models provide a possibility for calculating theoretical measurements such as brightness temperatures and wet path delays using the profiles. Figure 1-1. Conceptual diagram of the cascade process (left) to the brightness temperature (right) by RT model: The RT model calculates the brightness temperature using meteorological profiles provided by NWM 8 In order to find an improved method of retrieval, the investigations will be focused on the zenith wet delay (ZWD) by readjustment of the WVR measurements based on the model calculations. It may be possible to find the best readjustment through comparative investigations of geodetic parameters. This is similar to an adjustment process of the wet path delays, periodically fitting the WVR measurements to the radiosonde-derived wet path delays. In this study, radiosonde measurements will be replaced by the model calculations for the same purpose. As a final step, the readjusted ZWDs will be applied to geodetic VLBI sessions and their effects will be investigated mainly in terms of baseline repeatability and height precision. The key steps of this study are summarised below. 0. Calculate theoretical measurements using an RT model (MonoRTM; Clough et al., 1989, 2005) introducing profiles from ECMWF 1. Readjust the measured ZWDs based on the theoretical calculations by the models. 2. Apply the readjusted wet path delays to the geodetic VLBI sessions 3. Investigate the effects of the delay corrections to the geodetic VLBI sessions mainly in terms of baseline repeatability and height precision. When embarking on this thesis, it was expected that several geodetic VLBI sessions with useful WVR measurements at Effelsberg would be available eventually. However, in the course of time it turned out that only five sessions could be successfully observed. For this reason, the conclusions will have to be based on these five sessions alone. 1.2 Water vapour in Space Geodesy Water vapour is one of the most pending obstacles for ground based space geodetic observations. In particular, space geodetic technologies using microwave frequencies such as GPS and VLBI mostly suffer from uncalibrated water vapour-induced delays. However, its distribution in the atmosphere is impossible to accurately predict only with atmospheric models. Thus, the wet path delay is still the limiting factor for further improvements in space geodesy. As engineering technologies become more and more advanced, space geodesy is jumping to new levels of performance with unprecedented accuracy and precision particularly in positioning of global scale which is necessary for the maintenance of the Terrestrial Reference Frame (TRF). Until now, most of space geodetic networks are deployed on the surface of the Earth and use microwave signals for measuring distances and angles. The atmospheric water vapour contributes different excess path lengths to the distance measurements from individual sites in the global network, because each site is subject to different weather conditions. To meet the high performance of the observations themselves, the wet path delays have to be handled properly and corrected in more sophisticated manners Path delays induced by atmosphere Ground-based space geodetic instruments observe microwave signals that are emitted from sources in outer space, such as Quasars and artificial satellites. The 9 microwave signal inevitably passes through the atmosphere to be detected at the surface of the Earth. Figure 1-2 depicts a typical condition of geodetic VLBI observations with two distant telescopes. The microwave signal reaching ground receivers includes the overall effects of error sources between the sources and the receivers. The signal experiences slowing and bending while it passes through the atmosphere. The atmosphere can be divided into two layers that are descri
Related Search
Similar documents
View more...
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks